OnQ Blog

Putting camera and AI at the network edge

Dec 8, 2020

Qualcomm products mentioned within this post are offered by Qualcomm Technologies, Inc. and/or its subsidiaries.

If you're designing a consumer device like a home security system or an industrial device like a dash camera, you don't want to reinvent the wheel any more than you have to. Expensive, vertical products already exist for those use cases, so you're trying to design new products around affordable, mobile architectures and standards.

Based on the Qualcomm® QCS610 system-on-chip (SoC), the IPC610 Open Development Kit is now available from Altek. This reference design combines a camera, dual image signal processors (ISPs), video capture and playback at up to 4K (30 fps), artificial intelligence, audio codecs, hardware-based security and connectivity. It's geared for original design manufacturers (ODMs) and application developers who know mobile and want to avoid integrating multiple subsystems from scratch.

With the IPC610 kit, you can develop your roadmap into the Artificial Intelligence of Things (AIoT), where AI takes advantage of the growing number of IoT devices. Sample use cases include automotive dash cameras, consumer IP cameras for home security and enterprise IP cameras for monitoring premises to name just a few.

Go ahead. Run multiple machine learning models simultaneously.

The QCS610 SoC is a member of the Qualcomm Technologies, Inc. (QTI) Visual Intelligence Platform, a series of chipsets ideal for smart-camera products in the consumer and enterprise IoT spaces. Each member of the platform is highly integrated — power management, audio, image processing, connectivity — designed to reduce your bill-of-materials costs and help shorten your time to commercialization.

At the same time, the QCS610 offers the heterogeneous computing power needed for high-quality images and video with low power consumption. The result is 50 percent higher performance (compared to previous generations of QTI SoCs) and the power to run multiple machine learning models simultaneously, as shown in the video on our product page.

In that demo, three models are running in parallel on the same SoC:

MobileNetV2 SSD for detection — to indicate that a person is in the field of vision

DeepLab v3 for image segmentation — to separate the object in the foreground of the image from the background

PoseNet for pose estimation — to track physical movements in the field of vision

What does that mean for computing at the network edge? It means that you can execute more workloads locally on devices powered by the QCS610 instead of shuttling tasks to and from the cloud. That opens up opportunities for sample applications like face recognition, face/body detection, object classification and license plate recognition.

Combining hardware and software

The main hardware features of the development kit include the QCS610 SoC, camera sensors and a lens offering a 110- to 170-degree field of view. The kit supports 4G, Ethernet/RGMII and Bluetooth 5.x/Wi-Fi with antenna. (Support for 5G is coming soon.) Also built in is the Qualcomm AI Engine, dedicated hardware and software designed to accelerate on-device AI.

Supporting software includes:

Qualcomm® Neural Processing Engine SDK — For developing and running custom AI models based on industry-standard frameworks, then running them on the compute cores best suited to the workload

Qualcomm® Connected Camera SDK — An audio-visual framework with APIs for encoding multiple channels of AV from the camera, with pre- and post-processing

Open Dev Kit design — A turn-key approach to faster commercialization of consumer and enterprise hardware products

You'll find that the whole is greater than the sum of the parts when you can take advantage of AI inferencing at the network edge. You can also offer your customers a product that addresses the main objections to cloud-based AI, such as latency, privacy, connection and cost.

Get started

The IPC610 Open Development Kit is designed to help make top-tier camera technology affordable for mid-tier camera segments. It wraps up in a small package several AI/machine learning features formerly available only to high-end devices, then adds the heterogeneous computing and connection options needed at the network edge.

You can buy the kit now and get started right away on prototypes of your new smart-camera products.

Qualcomm QCS610, Qualcomm AI, Qualcomm Neural Processing SDK, and Qualcomm Connected Camera SDK are products of Qualcomm Technologies, Inc. and/or its subsidiaries.

Engage with us on

and

Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries ("Qualcomm"). Qualcomm products mentioned within this post are offered by Qualcomm Technologies, Inc. and/or its subsidiaries. The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.

Ravi Balasubramanian

Staff Mgr. Business Development, Qualcomm Technologies

Related News

©2021 Qualcomm Technologies, Inc. and/or its affiliated companies.

References to "Qualcomm" may mean Qualcomm Incorporated, or subsidiaries or business units within the Qualcomm corporate structure, as applicable.

Qualcomm Incorporated includes Qualcomm's licensing business, QTL, and the vast majority of its patent portfolio. Qualcomm Technologies, Inc., a wholly-owned subsidiary of Qualcomm Incorporated, operates, along with its subsidiaries, substantially all of Qualcomm's engineering, research and development functions, and substantially all of its products and services businesses. Qualcomm products referenced on this page are products of Qualcomm Technologies, Inc. and/or its subsidiaries.

Materials that are as of a specific date, including but not limited to press releases, presentations, blog posts and webcasts, may have been superseded by subsequent events or disclosures.

Nothing in these materials is an offer to sell any of the components or devices referenced herein.